Data driven methods to discover, stable, linear models of the helicity injectors on HIT-SIU from experimental data

The HIT-SIU experiment at the University of Washington continued the study of Steady Inductive Helicity Injection (SIHI) for the formation and sustainment of spheromaks. SIHI utilizes one or more injectors, which are each inductively driven by independent “voltage” and “flux” circuits. To better understand and control the behavior of these circuits on the HIT-SIU experiment, data driven methods, such as the Sparse Identification of Nonlinear Dynamics (SINDy)1, and the Dynamic Mode Decomposition (DMD), were implemented on current and voltage data from the injector circuits. The goal of these methods is to learn stable, linear models of the interaction of the injector circuits with the plasma. These models are then paired with optimal linear state estimation, and control to pair for linear quadratic Gaussian (LQG) control and estimation. Work supported by ARPA-E award DE-AR0001266.

1- SL Brunton et al. Proceedings of the national academy of sciences 113, 3932-3937 (2016)